10.5k views12 Upvotes59 Comments

Senior Director in Healthcare and Biotech, 1,001 - 5,000 employees
2 big concerns- first finding and retaining the talent to program it, and second is the building of inherent bias into the models that does not get picked up until very late into testing or post go-live.
Solutions Architect in Software, 501 - 1,000 employees
The right dataset is always the 98% to success. The more comprehensive, integral and noise-free the data is the less concerns you have with the model.

VP of Engineering in Software, 11 - 50 employees
My biggest concern about AI is the ethics (or lack of) behind it. AI projects should go through board review before being allowed to start and should go through regular checks until they are retired. They are powerful tools that can't be left in the wrong hands or unsupervised.
Chief Technology Officer in Healthcare and Biotech, 1,001 - 5,000 employees
Bias and ethics … in both cases it comes down to humans applying the technology appropriately and ensuring the training data is appropriate:
Chief Technology Officer in Media, 2 - 10 employees
My biggest concern about AI is what happens when it will be getting it to wrong hands.
Then would be the cost, no creativity, there would be more unemployment, emotionless and goes on. I am not on the side of 'Say no to AI' but these all cons comes to mind whenever the topic of AI floats around.
Sr. Director of Engineering in Software, 51 - 200 employees
The degree of intelligence in AI comes with quality of data used for training and its very rare to get perfect dataset in any field. So, the bais introduced due to this can lead to different results leading to difficult real life situations. So, applications of AI into fields that deal with human lives, economy, physical security, privacy like healthcare, banking, threat scanners etc. can be very concerning. The interpretation of a situation to get right context and handling within moral/ethical domain is still a big challenge with using AI. 
VP of Engineering in Finance (non-banking), 10,001+ employees
Primarily two things - 1) AI is built on probabilistic models that can be altered through the learning process that can be altered via data sets that can be modified. This means someone with bad intent can misuse it. 2) AI cannot think. It can simply process information that can result in biased results. Human life is not black or white. Our decisions are not black or white. 
Chief Technology Officer in Software, 11 - 50 employees
My main concern with supervised learning is biased training of the bots. Having neutral datasets are incredibly difficult to vet and maintain. 
CIO in Energy and Utilities, 1,001 - 5,000 employees
When AI is used in ways that are ethically questionable in order to achieve a certain outcome. It's simply a means to an end. 
Computer Science Lecturer in Education, 51 - 200 employees
The most concerns about AI are related to:

Determining the right data set, Data security and storage, Infrastructure and Integration

Content you might like